More than 100,000 patients are hospitalized in the United States each year for complications from ulcers induced by the use of nonsteroidal anti-inflammatory drugs (NSAIDs). Persons who use NSAIDs and have a history of ulcers are at highest risk for developing another ulcer. Cyclooxygenase-2 (COX-2) selective NSAIDs are becoming widely used in the treatment of arthritis in patients at high risk for ulcer complications. Chan and colleagues compared the COX-2 selective NSAID celecoxib with the combination of a standard NSAID (diclofenac) and a proton pump inhibitor (omeprazole) in the treatment of arthritis in patients with a recent bleeding ulcer.

The authors screened 396 patients taking NSAIDs for arthritis who presented with endoscopically confirmed bleeding ulcers; 287 patients (72 percent) were included in the analysis. The most common reasons for exclusion were absence of indication for prolonged NSAID use, renal failure, cancer, and declined consent. All enrolled patients tested negative for Helicobacter pylori infection or had successful eradication of infection according to gastric biopsy.

After follow-up endoscopy confirmed that their ulcers had healed, participants were randomized to treatment with celecoxib in a dosage of 200 mg twice daily plus placebo once daily, or diclofenac in a dosage of 75 mg twice daily plus omeprazole in a dosage of 20 mg daily for six months. Double-blinding was achieved by repackaging the medications in identical capsules. Patients were followed for recurrent ulcer, defined as hematemesis, melena, or a hemoglobin decrease of at least 2 g per dL (20 g per L) with an endoscopically verified recurrent ulcer.

Discontinuation rates were similar between treatment groups (13.3 percent for celecoxib versus 11.9 percent for diclofenac-omeprazole). Recurrent ulcer bleeding occurred in seven of 144 celecoxib patients (4.9 percent) and in nine of 143 patients taking diclofenac-omeprazole (6.3 percent).

Nonulcer adverse events were common in both treatment groups. Gastrointestinal problems occurred more frequently in the group treated with celecoxib (16.0 percent) than in the group treated with diclofenac-omeprazole (9.8 percent), but medication discontinuance rates were similar in the two groups (4.2 percent for celecoxib, 3.5 percent for diclofenac-omeprazole). Adverse renal effects (e.g., hypertension, peripheral edema, creatinine level greater than 2.2 mg per dL [200 mmol per L]) occurred in a substantial portion of both groups (24.3 percent of celecoxib patients and 30.8 percent of diclofenac-omeprazole patients).

The authors conclude that treatment with celecoxib or a combination of diclofenac and omeprazole had equivalent risks for recurrent ulcer bleeding in patients with arthritis who had a recent history of ulcer.